基于SQI的超分辨率图像重建算法

提出了一种基于自商图像(Self Quotient Image-SQI)的超分辨率图像重建算法.该方法首先利用SQI提取光照不变量作为图像特征,并假设光反射分量具有分段平滑的特性,近似认为每一个小的图像块具有相同的增益系数;然后在流形学习的框架下,借助局部线性嵌套的思想构建高分辨率图像和低分辨率图像块间的关系,从而实现了超分辨率重建和图像增强.仿真结果表明,该算法有效地克服了传统方法受光照因素影响的缺点,在提高分辨率的同时克服了光照因素的影响,特别是对阴影效应的消除具有明显效果.

A SQI-Based super-resolution algorithm

In this paper,a novel Self Quotient Image(SQI)-based super-resolution algorithm is proposed.In this method,First SQI is used to extract intrinsic,illumination invariant features of the image.Meanwhile,it is assumed that the reflectance is piece-wise smoothing and each patch has the same gain coefficient.Then the high resolution image is reconstructed based on the assumption that small SQI patches in low resolution space and patches in high resolution space share the similar local manifold structure.Experimental results demonstrate that the proposed method simultaneously achieves single-image super-resolution and image enhancement especially shadow removing.